On 2017/10/14 0:59, Weichao Guo wrote:
> Punching hole with a length which far larger than file size will
> cause a long time looping in truncate_hole after lock_op, it can
> be reproduced as the following:
> 
>         $ echo "abc" > file
>         $ xfs_io -c "fpunch 0 HUGE_LENGTH" file
> 
> This may cause other I/Os blocked a long time. Let's fix this case
> by setting the hole to end after the page that contains i_size.
You know, we can reserve blocks across EOF through fallocate(fd, 
FALLOC_FL_KEEP_SIZE,
ofs, len), so this patch will cause f2fs losing ability of punching preallocated
blocks across EOF.

How about below change:

diff --git a/fs/f2fs/file.c b/fs/f2fs/file.c
index cfee75bf88d9..91fd2ef7221b 100644
--- a/fs/f2fs/file.c
+++ b/fs/f2fs/file.c
@@ -847,7 +847,7 @@ int truncate_hole(struct inode *inode, pgoff_t pg_start, 
pgoff_t pg_end)
                err = get_dnode_of_data(&dn, pg_start, LOOKUP_NODE);
                if (err) {
                        if (err == -ENOENT) {
-                               pg_start++;
+                               pg_start = get_next_page_offset(&dn, pg_start);
                                continue;
                        }
                        return err;

Thanks,

> 
> Signed-off-by: Weichao Guo <[email protected]>
> ---
>  fs/f2fs/file.c | 8 ++++++++
>  1 file changed, 8 insertions(+)
> 
> diff --git a/fs/f2fs/file.c b/fs/f2fs/file.c
> index 517e112..1b44d26 100644
> --- a/fs/f2fs/file.c
> +++ b/fs/f2fs/file.c
> @@ -873,6 +873,14 @@ static int punch_hole(struct inode *inode, loff_t 
> offset, loff_t len)
>       if (ret)
>               return ret;
>  
> +     /*
> +      * If the hole extends beyond i_size, set the hole
> +      * to end after the page that contains i_size
> +      */
> +     if (offset + len > inode->i_size)
> +             len = inode->i_size + PAGE_CACHE_SIZE -
> +                     (inode->i_size & (PAGE_CACHE_SIZE - 1)) - offset;
> +
>       pg_start = ((unsigned long long) offset) >> PAGE_SHIFT;
>       pg_end = ((unsigned long long) offset + len) >> PAGE_SHIFT;
>  
> 

------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
Linux-f2fs-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel

Reply via email to